Most Indian ecommerce brands trying to grow sales start by changing their ads. Usually the ads are not the constraint. Working through this in the right order saves months and a great deal of money.
Fix the leaks before you increase spend
Paid traffic multiplies whatever your store already does. If the store converts at 0.8%, more traffic just loses money faster. Four leaks cost more than any ad optimisation can recover.
LeakTypical impactFix
Slow mobile load20–40% of sessions lostCompress images, cut apps, under 3s
Broken or missing trackingAlgorithm optimises blindServer-side CAPI + GA4 events
No trust signalsCart abandonment above 75%Reviews, returns policy, COD clarity
Checkout friction10–25% drop at paymentGuest checkout, UPI, fewer fields
In India specifically: if you do not offer UPI at checkout and are not clear about COD availability, you are losing a meaningful share of buyers before the ad has any chance to work. This is not an optimisation — it is table stakes.
The maths you need before spending
Almost every failed ecommerce ad account we audit has never calculated its allowable acquisition cost. Do this first.
1Average order value. Take the last 90 days, not your best month.
2Gross margin per order. AOV minus product cost, shipping, packaging, payment gateway and expected returns. Indian D2C brands routinely forget RTO cost, which on COD orders can be 15–30%.
3Repeat rate. What share order again within 12 months, and how many times.
4Allowable CAC. If margin per order is ₹450 and a customer orders 2.2 times a year, first-order contribution supports roughly ₹450 while lifetime supports closer to ₹990. Decide which you are willing to fund.
5Break-even ROAS. Divide AOV by gross margin per order. If AOV is ₹1,500 and margin is ₹450, break-even ROAS is 3.33. Anyone quoting you a "good ROAS" without knowing this number is guessing.
Campaign structure that works now
Both platforms have consolidated around automation. Fighting that with granular segmentation is the most common and most expensive mistake in Indian ad accounts.
Meta
—One prospecting campaign. Advantage+ Shopping or broad targeting, optimised for purchase. Not link clicks, not add-to-cart, unless volume genuinely prevents purchase optimisation.
—Six to ten distinct creatives. Genuinely different concepts across static, video, carousel and vertical Reels — not ten variations of one headline.
—One retargeting campaign. Site visitors, add-to-carts and engagers, excluding purchasers from the last 30 days. Different creative angle, not the same ads.
—Catalogue connected. Dynamic product ads consistently outperform static creative for retargeting in ecommerce.
—Leave placements automatic. Hand-picking placements restricts a system with better information than you have.
Google
—Performance Max with a clean feed. Feed quality — titles, images, attributes, GTINs — matters more than campaign settings. Most PMax underperformance is a feed problem.
—Brand Search kept separate. Always. Mixing brand and non-brand makes your reporting meaningless and inflates apparent ROAS.
—Standard Shopping alongside PMax for your top SKUs, so you retain visibility and control on what actually sells.
—Search campaigns on high-intent non-brand terms with tight negatives, mined weekly.
Budget split for Indian D2C
StageProspectingRetargetingBrand search
Launch (month 1–2)75%20%5%
Growing (month 3–6)65%25%10%
Scaled60%25%15%
Festive peak70%25%5%
The temptation is to over-invest in retargeting because its ROAS looks superb. That ROAS is largely borrowed — you are taking credit for demand prospecting created. Retargeting cannot grow a business on its own; it can only harvest more efficiently.
Scaling without breaking learning
1Raise budget by no more than 20–30% at a time, then wait 48–72 hours. Larger jumps reset learning and the resulting dip gets misread as failure.
2Scale the campaign, not by duplication. Duplicating a winning ad set fragments data and both copies underperform the original.
3Add creative before adding budget. Creative fatigue, not bidding, is what caps most accounts.
4Widen geography before narrowing audience. Adding tier-2 and tier-3 Indian cities usually costs less per acquisition than metro-only targeting.
5Watch blended CAC, not platform ROAS. Platform-reported numbers double-count. Blended — total spend divided by total new customers — is the only figure that survives scrutiny.
The festive question
Indian ecommerce concentrates a disproportionate share of annual revenue into roughly six weeks around Navratri, Dussehra and Diwali. CPMs rise sharply, sometimes 40–80% above baseline.
—Build audiences in the six weeks before. Cheap traffic then becomes cheap retargeting during peak.
—Accept a worse CAC during peak if lifetime value supports it. Volume compensates.
—Have creative ready in advance. Producing festive creative during festive season means launching late and paying more for it.
—Check inventory and courier capacity first. Selling what you cannot ship costs more in reviews and refunds than the revenue was worth.
WHERE THIS APPLIES
Ecommerce advertising reaches buyers across every Indian city, and tier-2 and tier-3 acquisition costs are frequently lower than metro rates. Careers Ninza teaches this live online, so learners join from Kolkata, Delhi NCR, Mumbai, Bengaluru, Hyderabad, Pune, Chennai, Ahmedabad, Jaipur, Chandigarh, Lucknow, Indore, Nagpur, Coimbatore, Asansol, Durgapur, Siliguri, Patna, Ranchi, Bhubaneswar, Guwahati, Gwalior, Noida and Gurugram. See all locations →
Market opportunity
Indian ecommerce continues to add buyers from beyond the metros, and much of that demand is served by brands with no serious performance marketing capability. The gap is not demand — it is operators who can calculate an allowable CAC, fix a conversion leak and read a blended number honestly. That skill is scarce relative to the number of brands that need it, which is why it commands both salaries and retainers.
Why Careers Ninza
—You run real budgets. Both our marketing programs end in a live campaign with tracked CAC, reviewed by a mentor — not a simulated exercise.
—Unit economics taught first. Allowable CAC, contribution margin and payback before any platform training, because the platform work is meaningless without them.
—Server-side tracking on the syllabus. CAPI, GA4 events and offline conversion upload, which is where most Indian accounts are genuinely broken.
—Practitioner mentors running live ecommerce budgets this quarter.
—No-Cost EMI on every program and twelve months of placement support.
Performance Marketing & Growth Hacking is three months at ₹24,999 on paid acquisition and CAC discipline. Ecompreneurship is six months at ₹49,999 and you launch an actual store during it.
SEE THE COURSES
